Hi all,
I have a Fujitsu SSD64GB with a SandForce SF-2241VB2-SPC controller that dropped to ROM mode after a spontaneous firmware panic. Goal is data recovery chip-off is not viable due to AES-128/DuraWrite/RAISE so I need to get production flashware running without FORMAT to keep the FTL tables intact.
Hardware details:

Controller: SF-2241VB2-SPC, HW revision 400026BB (reports hw version 0x0)
NAND: 8x Intel JS29F64G08ACMF3, 20nm MLC
Flash ID: 0x89, 0x88, 0x24, 0x4B, 0xA9, 0x84, 0x00, 0x00
Config ID: 28188 (confirmed correct by Genesis NAND ID match)
ROM mode via J3 service pin, presents as SandForce{200026BB} 306ABBR0, 64 sectors

I've been running Genesis v1.5 and v1.7 (fc14-32bit) on Fedora 14 32-bit live with libata.noacpi=1. MFP is FW_PPRO_5_8_2_Release.906190.release.359 (both CRC32 checksums pass). Both Genesis versions required binary patching to work the main issues were a 5-second sg_io_hdr timeout on the DOWNLOAD MICROCODE command in v1.7 (drive asserts BSY for 60-90 seconds during NAND programming), and various settle/identify timeouts that were too short.
The download itself now completes successfully and consistently Genesis reports "Successfully downloaded MFB to the drive" and the BSY hold confirms NAND programming is occurring. The problem is what happens after:
Without J3: drive enters an endless COMRESET cascade after the download. PHY is present throughout (kernel logs "link is slow to respond" rather than "link down") but SATA negotiation never completes. I tested with the AHCI controller fully unbound for 20 minutes to eliminate all host interference — same result. Drive never presents a valid IDENTIFY response.
With J3: drive always returns cleanly to ROM mode (64 sectors), every time.
This pattern is consistent across Feature 1 (monolithic 284KB transfer) and Feature 3 (64-sector chunked transfer). The drive has received 15+ successful MFB writes and the BSY hold has been getting progressively longer, which suggests NAND wear in the service area.
Has anyone seen this specific boot loop on SF-2241 where the download completes but the drive can't boot from NAND? I'm wondering whether there's a service area preparation step required before the MFB can load, or whether direct NAND programming of the boot area is viable if the controller genuinely can't recover via ROM mode.
Any advice appreciated.
Thanks

Brazen

_________________
A backup a day keeps DR away.